aikesommer.name - hello again?!?!

Awesome and legendary online pokertimer

aike — Mon, 22 Mar 2010 18:19:24 +0000

Every now and then we get together and play some Texas Holdem. Having one of those nights coming up and a slow sunday afternoon i decided to code a little blinds-timer to ease keeping track of time and blinds.

Features of my pokertimer

There is not much to list here, it lets you enter blind-levels and a duration. It will then count down and ring and flash when the time expired. The interface contains only the necessary information and is optimized for fullscreen use. In most web-browsers you can hit F11 to get a real fullscreen window.

To distract as little as possible you can use the space-bar to start and stop the timer, which should save you at least 3 minutes of mouse moving and clicking on the night!

Requirements for my pokertimer

Well, none basically. It works in any recent web-browser, i tested in Firefox and Chrome. So, just make sure you're not still using IE6 and have an online connection for the night.

Where to get my pokertimer?

Just head over to aike.me/pokertimer. It's completely free and ad-free. And beware, the numbers will be big!

Another heading including "my pokertimer"?

Yes, absolutely!

I hope you enjoy it. If you happen to have any questions or feedback, you can find my mail-address on the right.

New releases for AuthenticRoast and ViewNControl

aike — Thu, 04 Mar 2010 15:25:40 +0000

Today i fixed up 2 releases: AuthenticRoast 0.3.2 and ViewNControl 0.2.0

Both of them are feature releases, AuthenticRoast can now manage the authentication of users across contexts, so you can now have one webapp authenticate the user for another webapp running in the same container.

ViewNControl 0.2.0 adds basic mouse and keyboard input, which makes it actually almost usable! ;-)

New release of AuthenticRoast - Moved to Google code

aike — Wed, 03 Feb 2010 12:27:10 +0000

Over the last few months a few features and bug-fixes have piled up for AuthenticRoast, so i decided it's finally time for a new release.

The project has also been moved to Google code, and can be found here.

Other than some API-cleanup and bug-fixing the 2 most important changes are:

You can now register your authenticator module using a context-param.
The same authenticator instance will be used throughout the web-app lifecycle, so make sure it's thread-safe!

Let me know if you are having any problems with this release.
:-)

ViewNControl: VNC connections with pure HTML / JavaScript

aike — Fri, 25 Sep 2009 15:39:03 +0000

A while back i worked on a way to access VNC-sessions without requiring anything but a recent browser. This was just a proof-of-concept for a project that never went live. But it does work pretty good without bigger delays.

Today i finally went ahead and created a project on Google code for it. If you like check it out: http://code.google.com/p/viewncontrol/

I have tested it with x11vnc, no idea how far you can get with other VNC-servers. There is also no input at this time, so the usefulness is agueable. ;-) I have set the connection to be shared in x11vnc, and used a regular VNC-client for input during development.

If you like the idea and would like to help me push this in some useful direction, drop me a mail. I have also created a mailing-list at http://groups.google.com/group/viewncontrol-devel

Make that mouse-pointer stay out of invisible areas

aike — Thu, 24 Sep 2009 09:42:54 +0000

My last post to the planet is quite some time back and unfortunately i havent found the time i would have liked to work on multihead in KDE! I have tried to chase down and close a bug here and there, but have not done anything noteworthy!

Today i just wanna tell you about a tiny app i wrote. I have been using this for a few months now and cant live without it! ;-) So i figured others might enjoy it, too.

I simply called it "NoEnter" and all it does is make your mouse-pointer not go into those invisible areas on the framebuffer not covered by any of your screen. (This makes it totally useless, if you have only 1 screen or both running at the same resolution) I have searched for alternative ways to accomplish that (I figured, there needs to be a config-option for that) but havent found any, so i just hacked this up.

It not only helps to not "loose" the pointer, but also to not overshoot the panel, if it resides at the bottom of your smaller screen (as i have it).

The only requirement other than a recent KDE (with -dev/-devel packages) is the XTEST extension. This should be on by default for mostly anybody (AFAIK), but obviously this means: works only on X11, not on other KDE-platforms.

Now, wanna try it out? Here we go:

svn co svn://anonsvn.kde.org/home/kde/trunk/playground/utils/noenter
cd noenter
cmake .
make
sudo make install

If this ran without errors, you can just type "noenter" to start it. If the regions are miscalculated (never happened to me) you might not be able to use your mouse anymore, so be sure to have no important unsaved documents open, when you try for the first time! ;-)

Hope it helps somebody!
Aike :-)

RESTful web-services in Java using JAX-RS - Part 1: Getting up and running

aike — Mon, 14 Sep 2009 12:22:41 +0000

After having spent some time working with Suns JAX-RS implementation Jersey and integrating RESTful-services in a few products i work on, i thought it might be a good idea to share my experiences.

So i decided to write a few blog-posts. I am not sure yet how many parts there will be, my current plan are 3. In this first part i want to talk about the "trivial" things. It took me some time to gather all the information i needed, especially the Maven artifacts to include were not obvious to me.

Maven: Getting Jersey into your project

In this post i will assume, that you are working on a web-project and use Maven as build-tool. I run most of my apps with an embedded Tomcat, but you can just as well deploy the resulting war to any other container. Further i will explain how to use Suns Jersey, so if you have reasons to use a different JAX-RS-implementation, this part will not apply to you. (If you think these reasons might apply to others as well, please let me know, i am always happy to learn about alternatives.)

Last time i checked, the current versions of Jersey were not available on Mavens central repository. To get them add the following repository to your pom.xml (unless you already have for other dependencies):

    <repositories>
        <repository>
            <id>maven2-repository.dev.java.net</id>
            <name>Java.net Repository for Maven</name>
            <url>http://download.java.net/maven/2/</url>
        </repository>
    </repositories>

Now add the following artifacts:

    <dependency>
        <groupId>com.sun.jersey</groupId>
        <artifactId>jersey-server</artifactId>
        <version>1.0.3.1</version>
        <scope>compile</scope>
    </dependency>
    <dependency>
        <groupId>com.sun.xml.bind</groupId>
        <artifactId>jaxb-impl</artifactId>
        <version>2.1.12</version>
        <scope>compile</scope>
    </dependency>

This should pull in everything else to get a basic web-service up and running.

If you want to put some classes for your web-service into an external pojo, just add the following artifacts to its pom.xml:

    <dependency>
        <groupId>javax.ws.rs</groupId>
        <artifactId>jsr311-api</artifactId>
        <version>1.0</version>
    </dependency>
    <dependency>
        <groupId>javax.xml.bind</groupId>
        <artifactId>jaxb-api</artifactId>
        <version>2.1</version>
        <scope>provided</scope>
    </dependency>

Integration work: How to get Jersey running

First we need to map some requests to Jersey, this works by registering a servlet in your web.xml. Just add this:

    <servlet>
        <servlet-name>Jersey Servlet</servlet-name>
        <servlet-class>com.sun.jersey.spi.container.servlet.ServletContainer</servlet-class>
        <init-param>
            <param-name>com.sun.jersey.config.property.packages</param-name>
            <param-value>my.package.api</param-value>
        </init-param>
    </servlet>
    <servlet-mapping>
        <servlet-name>Jersey Servlet</servlet-name>
        <url-pattern>/api/*</url-pattern>
    </servlet-mapping>

Replace my.package.api with the real package you intend to use for your resource-classes.
We're mostly done with the implementation specific parts. The next steps involve only JAX-RS things, which lets you take your webservice-classes to any JAX-RS-implementation you like.

Hello World: Your first REST resource

Just really quick: in RESTful web-services everything is modeled as a resource, so basically you create operations for CRUD (create, read, update, delete). For different operations, different HTTP-methods are used, mainly GET, POST and DELETE. You might also use PUT sometimes. Important: it should never matter wether the same GET-request is sent multiple times.
Lets create a first, really basic resource-class:

    @Path("/notes")
    @Produces({"application/xml"})
    public class NotesHandler {
 
        @GET
        public String list() {
            return "Hello World, I still need some work to be useful!";
        }
 
    }

After deploying this to the container of your choice, you can open /api/notes in your web-browser. This should produce an xml-syntax-error, since only "Hello World, I still need some work to be useful!" is returned, which is obviously not valid xml.
So lets move on to create a class to wrap our result. We'll use JAXB-annotations to have this easily marshalled to xml:

    @XmlRootElement(name="notes")
    public class NotesList {
 
        private String result;
 
        public NotesList() {
        }
 
        public NotesList(String result) {
            this.result = result;
        }
 
        public String getResult() {
            return result;
        }
 
        public void setResult(String result) {
            this.result = result;
        }
 
    }

Now, also modify your handler-class like this:

        @GET
        public NotesList list() {
            return new NotesList("Hello World, I still need some work to be useful!");
        }

We now get valid xml, but this is still really useless. Lets add some logic to it:

    @Path("/notes")
    @Produces({"application/xml"})
    public class NotesHandler {
 
        /**
         * This is really, really wrong. Just so that we can do some playing around
         * without having to setup a database.
         */
        private static SortedMap<Long, Note> notes = new TreeMap<Long, Note>();
        static {
            add("Hello World, I still need some work to be useful!");
        }
 
        @GET
        public NotesList list() {
            NotesList result = new NotesList();
            result.getNotes().addAll(notes.values());
            return result;
        }
 
        private static Note add(String text) {
            Note note = new Note(notes.isEmpty() ? 1 : notes.lastKey(), text);
            notes.put(note.getId(), note);
            return note;
        }
 
    }
 
    @XmlRootElement(name="notes")
    public class NotesList {
 
        @XmlElement(name="note")
        private List<Note> notes = new ArrayList<Note>();
 
        public NotesList() {
        }
 
        @XmlTransient
        public List<Note> getNotes() {
            return notes;
        }
 
        public void setNotes(List<Note> notes) {
            this.notes = notes;
        }
 
    }
 
    @XmlRootElement(name="note")
    @XmlType
    public class Note {
 
        private Long id;
        private String text;
 
        public Note() {
        }
 
        public Note(Long id, String text) {
            this.id = id;
            this.text = text;
        }
 
        public Long getId() {
            return id;
        }
 
        public void setId(Long id) {
            this.id = id;
        }
 
        public String getText() {
            return text;
        }
 
        public void setText(String text) {
            this.text = text;
        }
 
    }

Enough with that GET, lets go POST something

Wouldnt it be great, if we could actually create notes, too? In RESTful web-services this is done with POST instead of GET, so we'll change the add method to handle POSTs:

        @POST
        @Path("/new")
        @Consumes("text/plain")
        public Note add(String text) {
            Note note = new Note(notes.isEmpty() ? 1 : notes.lastKey(), text);
            notes.put(note.getId(), note);
            return note;
        }

Also, remember to remove the static-block, that we added for testing.
You can now send a POST-request to /api/notes/new with the note as body to add a new note. I use the Firefox-extension Poster for this.

Okay... I think i'll just end this right here! I was going to write a little more, but this is getting really long as it is right now, so i'll just move the next steps to the next part!

I hope this is helpful for somebody to get started with JAX-RS.

SumtnSumtn goes public :-)

aike — Sun, 12 Jul 2009 10:01:11 +0000

We finally managed to get the last obstacles out of the way and go online!! Please check out our service at: http://sumtnsumtn.com/

There's not much else i have to say right now, have spent a little too much time in front of this screen for last weeks anyways!! ;-)

Problems with f:param inside h:outputLink

aike — Fri, 05 Jun 2009 16:59:22 +0000

There seems to be some weird issues with f:param inside h:outputLink. The first issue is "&" in param-values not being properly escaped, leading to completely wrong values on the receiver side.
The second seems a little more complex. If the value contains "--" (2 dashes) only the text of the link will be written to the page (no a-tag whatsoever).

This is on Suns JSF RI. If anybody knows anything about these issues i'd appreciate to hear about it!

:-)

Tomcat and UTF-8

aike — Fri, 01 May 2009 17:42:25 +0000

There's quite a few threads and pages about utf and tomcat out there, some mentioning URIEncoding in the cofig, others talking about an encoding-filter... None of which worked for me, entering special characters (in this case: umlaute) into a textarea gave me cryptic characters in my jsf-bean!

As it turns out, when the encoding-filter was invoked, it was already too late to change the request-encoding, the request object was already accessed in my very own authenticator-valve, making every later change useless!

After figuring this out, it was no problem to fix that, obviously! And the next AuthenticRoast release will also feature this... :-)

What i dont get is: Why is this not the default-behaviour with tomcat? Everything in java is (close to) unicode, hwy not there, too? Or at least provide a config-optin, that will work for POSTed data and not just GETs (as URIEncoding does)...

So much for now!!
:-)

Multihead in KDE 4.2.2

aike — Tue, 07 Apr 2009 10:35:24 +0000

Hello Planet!!

If everything goes well, this should be my first post, that appears on planet kde! :-) I've started hacking on kde for last years edition of gsoc and have since tried to get multihead better supported.
In this post i wanna talk about what works and what's still missing, hope you enjoy!! ;-)

What do i mean when i say multihead??

To prevent confusion: i mean xrandr/xinerama/mergedfb-style setups, not separate x-displays on different monitors! I dont think there is anybody actively working on the latter in kde 4...
Also, i'm mostly talking about adding/removing/changing screens on the fly, since "static" setups shouldnt cause any trouble for a while now!

What (should) work in KDE 4.2.2

(compared to earlier versions)

compositing continues to work after adding a monitor
plasma-panels which were added to monitors which were later disconnected, come back with the same size/applets/position when the monitor is plugged back in
maximizing windows after adding/removing a monitor results in the correct size, ie they dont go behind the panel and dont just maximize to half a screen only
probably some other things ;-)

What doesnt work

There's still lots of room for improvement, but nothing crucial, afaict. There is still no ui to configure kephal-configurations, but you can change your setup using xrandr or krandrtray.
Added or removed monitors can not be auto-detected, but there is nothing we can do about, it's more of a driver-problem!
If there's anything, that causes a real problem, please point me to the bug-report... And just maybe it might be able to resolve it!! ;-)

So much for now, have a nice sunny tuesday (at least it is here)!!
:-)